Process Reward Models (PRMs) trained using Single-Pass Annotation with Reference-Guided Evaluation (SPARE) methodology proposed in our AAAI-2026 paper
AI & ML interests
The Ubiquitous Knowledge Processing Lab researches natural language processing, text mining, eLearning, and digital humanities.
Recent Activity
View all activity
Organization Card
models
13
UKPLab/Qwen2.5-3b-spare-prm-math
Text Generation
•
3B
•
Updated
•
8
UKPLab/Llama-3-8b-spare-prm-math
Text Generation
•
8B
•
Updated
•
21
UKPLab/GritHopper-7B
Sentence Similarity
•
Updated
•
17
•
1
UKPLab/nlp4web-sciq-splade-bm25_negatives
67M
•
Updated
•
7
UKPLab/nlp4web-sciq-splade-random_negatives
67M
•
Updated
•
9
UKPLab/nlp4web-sciq-pointwise-reranker
67M
•
Updated
•
14
UKPLab/nlp4web-sciq-listwise-reranker
67M
•
Updated
•
12
UKPLab/nlp4web-sciq-dpr
66.4M
•
Updated
•
10
UKPLab/triple-encoders-dailydialog
Text Retrieval
•
Updated
•
12
•
6
UKPLab/agentbench-7b
Text Generation
•
Updated
•
12
datasets
15
UKPLab/sparp
Viewer
•
Updated
•
417k
•
123
•
1
UKPLab/HaluQuestQA
Viewer
•
Updated
•
698
•
17
UKPLab/PeerQA
Updated
•
358
•
3
UKPLab/m2qa
Viewer
•
Updated
•
21k
•
225
•
2
UKPLab/SLTrans
Viewer
•
Updated
•
6.98M
•
617
•
8
UKPLab/dapr
Viewer
•
Updated
•
124M
•
4.17k
•
2
UKPLab/DARA-Agentbench
Viewer
•
Updated
•
577
•
34
UKPLab/dara
Viewer
•
Updated
•
768
•
29
UKPLab/UKP_ASPECT
Viewer
•
Updated
•
3.6k
•
75
UKPLab/toxic_conversations
Viewer
•
Updated
•
1.8M
•
66
•
1